This is particularly in contrast to a study by Silbersweig et al. [52] reporting improved significance (greater signal to noise ratio and higher Z-score) of results obtained with 90 s compared to 60 s.
Unfortunately, haplotype analyses did not refine a smaller region of association as compared to that detected by single SNPs nor did haplotype analyses improve significance levels (data not shown); rather haplotypes turned out to be uninformative due to the larger number of haplotypic variants generated by the polymorphisms (including a microsatellite) typed across the gene.
Furthermore, we averaged ratios of duplicate spots on the array to improve significance.
With an Illumina spike experiment, we demonstrate dramatically improved significance of spiked genes over all relevant concentration ranges.
This demonstrates that our analysis generally agreed with more standard measures of epistasis and, in some cases, improved significance by reformulation in terms of directional model parameters.
Two other statistical approaches commonly used to improve significance estimates are a penalized t-test and a Z-test using intensity-dependent variance estimates.
The idea of normalization was initially offered as an ad hoc expedient to improve significance testing for differentially expressed genes in two-sample comparisons.
Pairs with hazard ratios (HR) showing a consistent direction between HEBCS and POSH data sets and improved significance in the pooled analysis were chosen.
Of note, although TYRP1/β-actin showed excellent correlation with TYRP1/S100B ρ=0.8888, P<0.001, Spearman's rho test) and OS (ρ=−279, P=0.005), it did not significantly correlate with DFS (P=0.123), supporting the fact that normalisation of TYRP1 with S100B improved significance.
Furthermore, expansion of the number of BRCA1 mutation carriers by including other family members in addition to the index cases in the study by Hollestelle and colleagues did also not improve significance, nor did the KRAS variant appear to modify breast cancer risk for BRCA1 mutation carriers [ 10].
